vvChevron Cracking" — A New Form of Hydrogen Cracking in Steel Weld Metals

Chevron cracking is a form of hydrogen induced cold cracking that can be eliminated through careful control of

the moisture content in basic type consumables

BY J. M. F. MOTA AND R. L. APPS

Introduction

Hydrogen induced cold cracking in the heat-affected zone has been one of the main problems in the welding of carbon manganese and low alloy steels. It has been gradually overcome by a better understanding of the phenomenon and its controlling parameters and, above all, by the introduction of better quality steels with lower carbon content. In fact, as the carbon content decreases towards 0.10% or less in the new steels, they tend to be free from the heat-affected zone cracking problem and the welding proce­dure is controlled by the need to avoid cracking in the weld metal (Ref. 1).

The relatively low carbon content and hardenability of most weld metals cre­ated a false sense of safety, i.e., that weld metals deposited with low hydrogen consumables would not be susceptible to hydrogen cracking. This feeling proved to be wrong in many cases. As the anti-hydrogen measures become less strin­gent, cracking has become more preva­lent in the weld metal. Some forms of cracking are undoubtedly due to hydro­gen while in other cases the mechanism is less certain.

One apparently new type of cracking was identified during the late 1960's and coincided with the then newly devel­oped basic agglomerated fluxes for sub­merged arc welding (SAW) (Ref. 2, 3). Due to the parallel pattern of the defects it was designated "chevron cracking," "transverse 45 deg cracking" or "stair­case cracking," but the first term is still

Paper presented at the 62nd AWS Annual Meeting held in Cleveland, Ohio, during April 6-10. 1981.

j M. F. MOTA is a Research Officer and R. L. APPS is Professor of Welding Technology, Cranfield Institute of Technology, Cranfield, Bedford. England.

the most used. The cracks are character­ized by their orientation, which is approx­imately transverse to the welding direc­tion and at 45 deg with the plane of the plates in a butt joint. The cracks are more easily identified in longitudinal sections of the weld metal cut perpendicular to the plane of the plates; in these sections some cracks intersect at 90 deg, produc­ing typical chevrons observable at low magnifications —Fig. 1. In the optical and scanning electron microscopes the stepped characteristic is readily apparent with two sets of components with the outline of a staircase —Fig. 2.

The problem has been found to occur in welds deposited by shielded metal arc welding (SMAW) and SAW, with a wide range of consumables. Stringent ultrason­ic inspection revealed the existence of chevron cracks in components such as heavy module structures for the North Sea oil production platforms, circumfer­ential seams in pressure vessels and restrained box sections in heavy engi­neering fabrications (Ref. 4).

The unusual orientation and morpholo­gy of the cracks (not easily related to other forms of weld metal cracking) cre­ated some speculation about the causes and mechanism. Their occurrence with

" low hydrogen" consumables (under conditions supposed to produce low weld metal hydrogen levels) tended to exclude this element as a possible cause of cracking. However, hydrogen was eventually assumed to be the most likely cause by excluding all the other possibili­ties on the available evidence. The work described here subsequently was in­tended to establish, on a firm basis, the controlling parameters and mechanism of chevron cracking.

Experimental Procedures

Cracking Test

It was known from previous works (Ref. 4, 5) that chevron cracks could be reproduced in a laboratory provided that a very rapid cooling rate was observed, particularly in the low temperature range (below the transformation temperatures). A special test based on the continuous water cooling of the base metal plates was specifically designed to produce a good control of the weld metal cooling rates and to investigate the conditions under which the chevron cracks occur (Ref. 6).

The base metal used in the tests con-

.

y . r ; •- •<•• - ' • • .

. . " a , ' .

K / ' S

p. * : W A

. • ? . • • • . £> • . . . "

\

Fig. I — Chevron cracking in a longitudinal section of an industrial weld. Nital etch; XI (reduced 26% on reproduction)

222-s | JULY 1982

Table 1—Chemical Compositio

Consumables (wire + flux)

SD3 + OP4 1TT

S4 + OP4 ITT SD3/M + OP41TT

SD2/3Ni + OP41TT

SD3/MO + OP41TT

SD3/INi ' /2Mo + OP41TT

SD2/ !Cr1Ni'/2Mo + OP41TT

n and Mechanical Properties

C

.11

.050

.14

.060

.055

.059

.08

Mn

1.24

1.37 1.40

1.05

1.13

1.33

.85

of Submerged Arc Welds

Chemical composition, wt —

Si

.15

.18

.18

.20

.18

.24

.19

S

.018

.010

.017

.010

.010

.05

.018

P

.021

.022

.019

.017

.016

.009

.022

Ni

.07

.03 1.50

2.94

.09

1.26

1.01

w

Cr

.03

.03

.07

.08

.06

.05

.84

Mo

.03

.02

.04

.13

.51

.63

.50

Cu

.16

.23

.23

.16

.16

.24

.22

Last run

222

221 246

249

251

266

277

Hardness HV5 Middle Refined

runs

203

207 225

226

223

257

270

regions

186

190 216

210

217

232

254

Mechanical YS,

N/mm 2

> 4 2 0 (> 370)

(b)

450-500 (390-440)

(b)

(410-440) 500-540

(460-500) (b)

(500-560) 650-700

(690-730)

properties'3' UTS,

N/mm2

580-630 (520-560)

(b)

560-630 (520-570)

(b)

(565-595) 630-680

(570-630) (b)

(600-680) 750-800

(800-830)

<a,Mechanical properties according to the consumable manufacturer tor weld metals in the as-welded condit ion: equivalent values for weld metals in the stress relieved condition are presented in brackets. YS -y i e l d strength; UTS-u l t imate tensile strength (ksi = N / m m 2 + 6.894757). <b,Not available

sisted of 38 m m (IV2 in.) thick and 450 m m (17.7 in.) long carbon manganese steel plates. The pre-set of the plates, pr ior to we ld ing, is shown in Fig. 3. N o external restraint was appl ied, and the pre-set of the we ld preparat ion was planned to ensure that the edges of the joint become parallel w h e n it was c o m ­pleted. The wate r f l o w was kept at 15 l i ters/min (4 gal lons/min), a l though it was found that the f l o w rate cou ld be varied w i th in w i d e limits w i thou t a significant effect on the we ld metal cool ing rates.

Af ter comple t ing each run, about 2 minutes (min) w e r e a l lowed be fo re remov ing the excess flux a n d / o r slag. The we ld metal coo led to the water temperature , 1 8 - 2 5 ° C (64-77°F) , shortly af terwards. The per iod b e t w e e n runs was kept constant at about 7 -10 min, but in some cases this was impossible and one or t w o breaks of Vl to 1 hour (h) occur red dur ing a test.

Fig. 2 — Details of a chevron crack: A — longitu­dinal section; nital etch X200. B — fracture sur­face, SEM X80 (A and B reduced 38% on reproduction)

After comple t ing each w e l d , at least 72 h elapsed be fo re sectioning the w e l d metal longitudinally to al low for any delayed cracking. Sectioning was carried out w i th a vertical band saw, and the base metal plates could then be used for subsequent testing after being dressed to the desired prof i le by hand grinding.

Welding Procedures

T w o series of tests w e r e carried out using the SAW and S M A W processes respectively. S A W was carr ied out w i th a commercial ly available basic agglomer­ated flux and a series of filler metal wires w i t h increasing alloying content (strength): SD3; S D 3 / 1 N i ; S D 2 / 3 N i ; S D 3 /

I/2 M o ; S D 3 / 1 N i ' / 2 M o and S D 2 / I C M N i ' / i M o (classification according to IIW recommendat ion) .

Table 1 shows the chemical compos i ­t ion of the welds and the typical mechan­ical propert ies according to the consum­able manufacturer . For each combinat ion the flux was used " bak ed for one hour at 4 5 0 ° C " (842CF), "as - rece ived" and " d a m p . " Typical moisture contents of the f l u x / w e l d metal hydrogen contents are shown in Table 2.

The w o r k also included S M A W using commercial ly available electrodes, con ­forming to A W S E8018 C 1 , and specially manufactured batches of basic, cellulosic and rutile electrodes in tended to match ,

Table 2—Moisture in the Submerged Arc Welding Flux and Corresponding Weld Metal Diffusible Hydrogen Levels

Condition of flux

B 450°C<a) AR<b» D ( 0

H 2 0, "o

0.020 0.033 0.052

Diffusible hydrogen, ml H2/100g weld metal

2.5 3.7 6.9

' • 'B 450 C - f l u x baked for 1 hour at 450 C (842 r) before being used (b)AR —"as received." ( c J D— "damp. " i.e.. flux exposed to the workshop atmosphere for 10 days wi th increased moisture content.

Fig. 3 - Weld preparation for the continuous water cooling test

W E L D I N G RESEARCH SUPPLEMENT I 223-s

Table 3

Test plate

TP1 TP2

TP3 TP4 TP5 TP6 TP7 TP8 TP9

TPIO TP1 1 TP12 TP13 TPI4

TP15 TP16 TP17 TP18 TP19 TP20 TP2I TP22 TP23

—Incidence of Cracking Identified by Magnetic

Consumables Condition (wire + flux) of flux

SD3 + OP4 ITT B450°C(a)

SD3 + OP4 ITT AR<b»

SD3 + OP41TT D(c)

S4 + OP41TT B450°C S4 + OP41TT D SD3/1Ni-FOP41TT B450°C SD3/1Ni + OP41TT AR SD3/1Ni + OP4ITT D SD2/3Ni + OP4 ITT B450°C

SD2/3Ni + OP41TT B450°C SD2/3Ni + OP41TT B450°C SD2/3Ni + OP41TT AR SD2/3Ni + OP41TT D SD2/3Ni + OP41TT D

SD3/Mo + OP4 ITT B450°C SD3/Mo -1- OP41TT AR SD3/Mo + OP4 ITT D SD3/1Ni'/2Mo + OP41TT B450°C SD3/1Ni'/2Mo + OP41TT AR SD3/1Ni'/2Mo + OP41TT D SD2/1Cr1Ni'/2Mo + OP41TT B450°C SD2/1Cr1Ni'/2Mo + OP41TT AR SD2/1Cr1Ni'/2Mo + OP41TT D

Particle Inspection

Max. no. of cracks > 1 mm in a

longitudinal section 100

45 deg

— 2

10

-13

-4

16

—

--10 45 25

-6

27

-----—

n m long Vertical

— —

------—

----

----12 25

-10 33

Continuous

Longest crack mm

— 2

4.5

-3.5

— 1 4.5

—

— — 2.5 4.5 6

— 3.5 7

-2 3

-8

10

Water Cooling Tests)

Average crack length,

mm

-2

3

2.5

-1 2.5

—

--1 3 3.5

-2 3.5

-1 2

-1.5 2.5

Notes

Some fissures < 1 mm (0.04 in.) were found

As above

As above

As above As above Tensile test specimens were machined

from this weld metal As above As above Some fissures < 1 mm (0.04 in.)

All runs were deposited in the same direction. Test monitored by acoustic emission

Some fissures < 1 mm (0.04 in.) As above

As above As above

As above As above

(a)B450 C - flux baked for 1 hour at 450 C (842 F) before being used. l b ,AR — "as-received." ( c ) D - " d a m p , " i.e., flux exposed to the workshop atmosphere for 10 days wi th increased moisture content.

as far as possible, the chemical composi­tion of the SA welds. Unfortunately, this was not always possible, and some of the compositions deviated considerably from those predicted. The basic electrodes were used in three conditions —"baked one hour at 400°C" (752°F), "as-received" and "damp"; the batches of cellulosic and rutile electrodes were used in the "as-received" condition.

The nominal heat input selected for both processes was 2.4 k)/mm (61 k)/ in.). The welding parameters were 500 A, 30 V; the welding speed was 380 m m / min (15 ipm) for SAW. SMAW was car­ried out at about 230 A, 24 V with a welding speed of 140 mm/min (5.5 ipm). The welding direction was reversed after each layer of 2 runs for SAW and 3 runs for SMAW.

Nondestructive Testing

After sectioning the strip of weld metal and backing bar, ultrasonic inspection was carried out to investigate the number and distribution of cracks for each test. This information was generally taken into account in further sectioning of the spec­imen prior to magnetic particle inspec­tion.

In order to allow for more precise macroscopic examination, each strip of weld metal was ground smooth with a surface planisher; surface or near surface

cracks were detected by magnetic parti­cle inspection. The type, size and number of cracks were recorded for each test.

Results

Macroscopic and Magnetic Partical Examination

About 60 tests were carried out using different consumables and varying the moisture content of the ones of the basic type. A typical set of results is illustrated in Table 3 for the SA welds. Although the precise type and number of cracks varied from one test to another, the results fall consistently within distinct groups deter­mined by the condition of the consum­ables and weld metal chemical composi­tion. For the purposes of this work, the weld metal cracks were classified into:

1. Microcracks or microfissures — cracks less than 1 mm (0.04 in.) long, generally oriented perpendicularly to the axis of the weld.

2. Chevron or 45 deg cracks-cracks transverse to the weiding direction which make an angle of 45 deg with the weld­ing axis when observed in a longitudinal section perpendicular to the plane of the plates.

3. Vertical or transverse cracks — cracks oriented perpendicularly to the welding axis.

Within the range of consumables examined, it was observed that, for

welds with consumables of the basic type (i.e., SAW basic agglomerated flux and SMAW electrodes with basic covering), the occurrence of cracking was primarily determined by the condition of the con­sumables in terms of moisture content. Macrocracks were consistently absent in welds with consumables baked at the maximum temperatures recommended by the manufacturers and kept in a hot quiver during welding; with other welds the degree of cracking increased with the moisture content of consumables. In the case of basic electrodes with very high moisture contents in the coating, cracking tended to assume the form of microfis­suring; porosity occurred often in these cases.

Welds made with cellulosic electrodes, which were not submitted to any treat­ment prior to welding, produced large numbers of microcracks as well as chev­ron cracks and transverse (vertical cracks). The welds made with rutile elec­trodes produced microcracks and porosi­ty, but no chevron cracks.

While cracking occurrence depended on the moisture contents of the basic electrodes and SAW fluxes, the type of macrocracking depended on the weld metal composition/strength. Chevron cracks were always associated with medium strength weld metals. When the strength was increased above a certain limit, the cracks were oriented vertically,

224-s | JULY 1982

Fig. 4 —Staggered non-linked microcracks. A —first stage; B — intermediate stage. Nital etch; X200

i.e., perpendicular to the welding direc­tion.

Metallographic Examination

Longitudinal sections from all speci­mens containing macrocracks were observed in the optical microscope with the objectives of establishing a relation between the morphology of the crack and the weld metal microstructure, and of comparing them with others in indus­trial samples. There were no major differ­ences between the cracks observed in the industrial samples and in the laborato­ry tests; consequently, the results and comments below refer to both except where otherwise stated.

Chevron cracks were found to occur in both the as-deposited and reheated weld metal, showing a characteristic stair­

case appearance with vertical and hori­zontal components —Fig. 2. The vertical components of the cracks in the as-deposited metal tended to occur pre­dominantly in the pro-eutectoid ferrite (at the prior austenite grain boundaries), but often they propagated away from it or followed a different path.

Many examples were found of non-linked microcracks with a special arrange­ment very similar to the vertical compo­nents of the chevron cracks, Fig. 4A. In some cases there was an indication that these microcracks had started to link, but the process was not completed —Fig. 4B. In the longitudinal horizontal plane, i.e., in an orientation parallel to the plane of the plates, chevron cracks were found to be almost transverse to the welding direc­tion and had the typical morphology illustrated in Fig. 5.

Vertical cracks (i.e., perpendicular to the axis of the weld or just slightly in­clined—Fig. 6) were typical of the higher strength weld metals used in the labora­tory tests. In some cases the cracks pre­sented a staircase morphology reminis­cent of the 45 deg cracks, but this was unusual. In general the cracks did not follow a typical path; they could be both inter- and transgranular, continuous or intermittent and sometimes branched.

Microcracks or microfissures such as shown in Fig. 7 could —with one excep­tion—be easily found in all weld metal cracking tests. The exception consisted of SA welds with the flux baked 1 h at 450°C (842°F). The microfissures were randomly distributed through the longitu­dinal sections, but their density was reduced in the top runs and in the runs containing high numbers of macrocracks. Microfissuring was the typical form of cracking in weld metals deposited with rutile electrodes; in some cases microfis­suring was found to be nucleated at a void. In general these weld metals did not develop macrocracks except for the most highly alloyed deposits.

Fractographic Examination

In the scanning electron microscope (SEM), at low magnifications, the staircase appearance of the chevron cracks was easily recognized — Fig. 2. After increasing the magnification, it was readily apparent that there were two distinct sets of com­ponents which could be observed in planes approximately perpendicular to each other. The vertical components pre­sented mainly quasi-cleavage (Fig. 8), but some showed relatively smooth areas — Fig. 9 (the significance of which is dis­cussed later in this paper). The origin of each vertical component was not always clear; in some cases, however, the com­ponents showed a pattern radiating from

-fv&VbiaizkiwHKL r

m!&

•

Fig. 5 —Chevron crack. Longitudinal section parallel to the plane of the plates. Nital etch; X200 (reduced 38% on reproduction)

Fig. 6— "Vertical crack" in a SAW deposit with Fig. 7—"Microfissures" in a SMAW deposit. SD3/1N 1/2 Mo wire. Nital etch; X80 (reduced Nital etch X20 (reduced 41% on reproduc-38",, on reproduction) tion)

WELDING RESEARCH SUPPLEMENT 1225-s

Fig. 8 - Chevron crack SEM: A - vertical component of a chevron crack; B — detail from A (reduced 41% on reproduction)

Fig. 9 — Chevron crack SEM: A — vertical component of a chevron crack with signs of exposure to high temperature; B — detail from A (reduced 41 % on reproduction)

the center where a void (pore or inclu­sion hole) could sometimes be found. The horizontal components showed essentially dimples —Fig. 10.

The fracture surface of the vertical cracks observed in the high strength weld metals consisted of areas of inter- and transgranular failure —Fig. 11. The mode of failure through the grains was quasi-cleavage. However, regions of dimples could be found linking areas of brittle failure.

Microfissures were observed for all weld metal compositions. The ones found in the medium strength weld met­als presented a mode of failure similar to the vertical components of the chevron cracks, but tended to be bigger in size. Fissures occurring in the runs below the top one very often presented smooth regions, resembling those found in the chevron cracks as referred to above.

Carbon extraction replicas from the fracture surface of some chevron cracks were observed in the transmission elec­tron microscope, after having examined the corresponding surfaces using the SEM. This work confirmed that the frac­ture surface of most cracks consisted of areas of quasi-cleavage alternating with areas of dimples with no evidence of thermal facets or grooves. These features were found, however, in a limited num­ber of vertical components of chevron cracks which showed smooth areas unre­solved in the SEM.

Discussion

When the present research program on chevron cracking was initiated, two different explanatory theories had just appeared (Ref. 4, 5). Insufficient evidence did not allow the controlling parameters of the phenomenon to be firmly estab­lished, leaving the field open to some controversy over the causes and mecha­nism. Keville's theory (Ref. 5) suggested that cracking resulted from a combination of hydrogen and stresses; this seemed to be supported by industrial evidence. However, the alternative presented by Tuliani could not simply be dismissed on the grounds that the features identified on the fracture surfaces (typical of expo­sure to high temperatures) were caused

-ry. Fig. 10 — Horizontal component of a chevron crack (reduced 38",, on reproduction)

by reheating on the deposition of subse­quent runs of weld metal (Ref. 7).

Existing evidence pointed towards hydrogen as being the main factor con­trolling the occurrence of chevron crack­ing. Nonetheless, it was possible that some form of hot cracking could be involved. This possibility was investigated and rejected through the study of a wide range of chemical compositions and car­rying out hot tensile tests on weld metals already studied for cracking susceptibility (Ref. 8). The influence of hydrogen on the cracking incidence was studied by using the SAW basic flux and the basic SMAW electrodes with controlled mois­ture contents, producing different hydro­gen contents in the weld metal. In brief, it was intended to support the metallo­graphic and fractographic evidence with extensive experimental results.

The cracking test designed by the authors was substantially more severe than the one ascribed to Tuliani (Ref. 4) and used in other works (Ref. 5, 9, 10). The use of continuous water cooling had the advantage of an easier control over the weld metal cooling rates. However, it was designed in a way that only the low temperature range of the weld metal cooling cycle (below the transformation temperatures) would be affected. In these circumstances the major effect would be on the retention of hydrogen.

Cracking tests results showed conclu­sively that hydrogen was the most impor­tant factor controlling the occurrence of chevron cracking and that the phenome­non must be due to a hydrogen induced cracking mechanism. This conclusion is based, primarily, on the fact that cracking can be entirely controlled by the mois­ture level in the consumables/hydrogen level in the weld metal. It was further substantiated by the fractographic exami­nation which established a good correla­tion between the features observed on the vertical components of the chevron cracks, microfissures and the areas of brittle failure on the surface of hydrogen charged tensile specimens in similar weld metals.

The comparison between the fracture surfaces of chevron cracks and hydrogen

Fig. II- "Vertical crack" in a SAW deposit made withSD3/1Ni '/2 Mo wire (reduced 38% on reproduction)

226-s | JULY 1982

induced cold cracks included also some cracks produced by the Y groove Tekken test*. Single runs of weld metals, depos­ited with SMAW electrodes which were known to produce welds prone to chev­ron cracking were studied (Ref. 8). The longitudinal cracks found in these tests often had a staircase morphology with two sets of components intersecting at an angle —Fig. 12; in the SEM they were found to be of the same nature of the vertical and horizontal components of the chevron cracks. This finding is partic­ularly important in showing that the stair­case pattern corresponds to a more gen­eral morphology of hydrogen cracking than has been realized.

The observation of quasi-cleavage in the vertical components of the chevron cracks, with sharp facets, is consistent with a cold cracking mechanism. Howev­er, the existence of smooth areas in some of them required close examination. It was proved, in subsidiary work by the authors, that these features could be closely reproduced by reheating chevron cracks through a convenient thermal cycle and could also be observed in multi-run Tekken tests. In these circumstances it is believed that the smooth areas in the chevron cracks, in subsurface runs, result from reheating and not from a hot cracking mechanism. This possibility was investigated further by Bihari (Ref. 8) who carried out hot tensile tests in the range 600-1320°C (1112-2408°F) for weld metals known to be prone to chevron cracking. However, no "ductility dip" was observed and, once again, the hot cracking mechanism as triggering chevron cracking was con­sidered unlikely.

The practical question has been answered, i.e., the weld metal hydrogen level is the controlling factor in the occur­rence of the types of cracking under consideration. However, the scientific interest remains: How do the chevron cracks occur? Why, in certain cases, do the cracks become oriented perpendicu­lar to the welding axis? Why does micro­fissuring become the preferential form of cracking in other cases?

To answer these questions, a mecha­nism has been advanced by the authors that integrates recent theories on "hydro­gen induced cracking."

A Mechanism for Chevron Cracking

Chevron cracks occur in carbon-man­ganese and low alloy steels by a two stage mechanism. The critical event is the occurrence of the vertical components due to "hydrogen embrittlement." The

"Tekken test: A weldability test of lapanese origin - Y groove looks like a tilted Lehigh groove specimen (Ref. 12).

Fig. 12 — Cross section of a longitudinal crack. Y groove Tekken test. Nital Nital etch; X5 (reduced 50% reproduction)

second stage consists in the linking of the vertical components due to the excessive stress between adjacent crack tips; this may also be assisted by hydrogen. A two stage cracking mechanism was initially suggested by the different cracking modes and orientation of the vertical and horizontal components when observed in the SEM. Further support came from the existence of non-linked staggered vertical microcracks and intermediate stages of linking alongside fully devel­oped chevron cracks.

In the conditions under which chevron cracking occurs, it is reasonable to assume that the transverse and through thickness residual stresses are small com­pared with the longitudinal stresses which may be of yield strength magnitude. Thus, it was expected that initial failure would occur essentially perpendicular to the maximum tensile stress. In fact, this was confirmed by the preferred orienta­tion of microcracks and vertical cracks (in the high strength weld metals), although the 45 deg cracks seem to be an excep­tion.

An important aspect of chevron crack­ing is its occurrence for relatively low levels of hydrogen. This fact may be understood in terms of the segregation of this element that may occur at grain and cell boundaries and form Cottrell atmo­spheres at dislocations; the critical level may be exceeded locally with extremely low overall concentrations (Ref. 11).

Based on the experimental results reported herein and the above argu­ments, the following model for the for­mation of chevron cracks is advanced:

1. Hydrogen has to be introduced in the weld metal above some critical con­centration depending on the microstruc­ture, the stress level and the diffusion coefficient of hydrogen.

2. Due to the longitudinal residual stresses in the weld metal, which are of yield strength magnitude, slip bands occur at approximately 45 deg with the welding direction.

3. Dislocations transport hydrogen along the slip bands until they find a barrier such as a grain boundary, a pre­existing crack or an inclusion where they

form pile-ups and concentrate hydrogen at a rate much faster than would be possible by hydrogen diffusion alone. Once the critical hydrogen concentration for the existing state of stress is reached, a microcrack occurs with an orientation perpendicular to the maximum tensile stress. This process is repeated along the slip bands producing a series of staggered microcracks along an orientation at 45 deg with the welding direction.

4. The microcracks are arrested when they enter a region where hydrogen falls below the critical level for the existing state of stress at the crack tip. The growth will not be restarted until hydro­gen is built up to the critical level or an increase in stress is observed.

5. When a microcrack is arrested, plastic slip at its tip may occur and this develops in slip band formation; further microcracks may occur along these slip bands as referred to in step 3.

6. The existence of high shear stress between the tips of adjacent microcracks generally results in rupture and linking of the staggered microcracks during the sec­ond stage of the cracking process, which may be assisted by hydrogen.

This model is based on the fact that hydrogen present in steel (once concen­trated above a critical level) would facili­tate the growth of crack nuclei. It does not specify the mechanism by which hydrogen embrittles the steel and is valid regardless of whichever embrittlement theory one cares to accept.

In the context of the above model, the existence of hot microcracks as crack initiators (Ref. 4) must be interpreted as a valid but probably infrequent or a suffi­cient but not necessary cause for a verti­cal component of a chevron crack to occur; however, hydrogen would be required for the propagation of such microcracks.

The occurrence of microfissuring as a preferred form of cracking when the weld metal hydrogen level is too high can now be understood. Above a certain hydrogen level, the microcracks may nucleate easily and do not depend on the transport of hydrogen by dislocations; in these circumstances the microcracks, instead of being staggered, occur ran­domly. If the hydrogen level is low, nucle­ation is very difficult; in this case, micro­cracks are simply eliminated or a few are randomly distributed.

The change in the orientation of the macrocracks from 45 deg to perpendicu­lar to the welding direction, for the high strength weld metals, is interpreted here­in in terms of an alteration in the cracking mechanism rather than a change in the intrinsic nature of the cracks. High strength weld metals are generally more susceptible to hydrogen embrittlement due to the metallurgical changes required to strengthen them. In this case the cracks may nucleate and propagate with lower hydrogen levels without relying on the

WELDING RESEARCH SUPPLEMENT I 227-s

transport of that element by dislocations along 45 deg slip bands. A second reason may be the relative weakening of the grain boundaries as the matrix is strength­ened, offering an easy, preferential path for the crack growth. In these circumstances the vertical cracks grow continuously through the material, although the process may be discontinu­ous in time as is typical of hydrogen cracking.

Conclusions

of 1. Chevron cracking is a type hydrogen induced cold cracking.

2. Chevron cracking can be elimi­nated simply by a careful control of the moisture content in the consumables of the basic type.

3. Microfissuring may become the preferential form of cracking, if the hydrogen level is too low or too high.

4. The orientation of the transverse, hydrogen induced cracks in the weld metal depends primarily on the weld metal composition/strength; chevron cracks are typical of medium strength weld metals.

5. A two-stage cracking mechanism was proposed to explain the occurrence and orientation of the chevron cracks.

References

1. Dolby, R. E. 1977. The weldability of low carbon structural steels. Welding Institute Research Bulletin 18(8):209-216.

2. Thomas, S. N. C. 1969. The implications of weld cracking in practice. Metal Construc­tion 7(2): 142.

3. Cotton, H. C. Ibid. 4. Tuliani, S. S. 1976. A metallographic study

of chevron cracks in submerged arc weld metals. Welding Research International 6(6):19-45.

5. Keville, B. R. 1976. An investigation to determine the mechanism involved in the formation and propagation of chevron cracks

in submerged arc weldments. Welding Research International 6(6):47-66.

6. Mota, |. M. F.; Apps, R. L; and ]ubb, ]. E. M. 1978. Chevron cracking in manual metal arc welding. Proc. Conf. on Trends in Steels and Consumables for Welding. Welding Insti­tute, paper no. 18.

7. Farrar, R. A. 1977. The nature of chevron cracking in submerged arc weld metals. Weld­ing Research International 7(2):85-88.

8. Bihari, L. M. 1976. Relation of hydrogen cracking and high temperature ductility to chevron cracking in steel weld metal. M.Sc. thesis, Cranfield Institute of Technology.

9. Wright, V. S., and Davison, I. T. 1978. Chevron cracking in submerged arc welds, Proc. Conf. on Trends in Steels and Consum­ables for Welding. Welding Institute, paper No. 38.

10. Chew, B., Private Communication. 11. Savage, W. F., et al. 1978. Hydrogen

induced cracking in HY-130 steel weldments. Welding lournal 57(4): 118-s to 126-s.

12. Vasudevan, R.; Stout, R. D.; and Pense, A. W. 1981. Hydrogen-assisted cracking in HSLA pipeline steels. Welding journal 60(9): 156-s.

WRC Bulletin 275 February, 1982

The Use of Quenched and Tempered 2 - l / 4 C r - l M o Steel for Thick Wall Reactor Vessels in Petroleum Refinery Processes: An Interpretive Review of 25 Years of Research and Application by W. E. Erwin and J. G. Kerr

This WRC Bulletin focuses on the evolution of modern hydroprocessing reactors fabricated f rom quenched and tempered 2 - l / 4 C r - l M o steel. It reviews and summarizes approximately 25 years of industry development and application of 2 - l / 4 C r - l M o steel in hydroprocessing reactors. Some discussion of ongoing research and needs for future research are identif ied.

Publication of this Bulletin was sponsored by the Subcommit tee on Thermal and Mechanical Effects of the Pressure Vessel Research Commit tee of the Welding Research Council.

The price of WRC Bulletin 275 is $13.00 per copy, plus $3.00 for postage and handling. Orders should be sent with payment to the Welding Research Council, 345 E. 47th St., New York, NY 10017.

WRC Bulletin 276 April, 1982

A Summary and Critical Evaluation of Stress Intensity Factor Solutions of Corner Cracks at the Edge of a Hole by R. L. Cloud and S. S. Palusamy

When the probable initial flaw size and the crack growth rate for a given material are known, crack growth behavior is calculable on the basis of loading and stress profile through the use of f racture mechanics and parameters. This report summarizes and critically evaluates Stress Intensity Factor (SIF) solutions to corner cracks at the edge of a hole.

Based on recommendations f rom this study, a compilation of SIF values for nozzle cracks is proposed.

Publication of this paper was sponsored by the Task Group on Nozzle Crack Growth of the Subcommittee on Reinforced Openings and External Loadings of the Pressure Vessel Research Commit tee of the Welding Research Council.

The price of WRC Bulletin 276 is $11.50 per copy, plus $3.00 for postage and handling. Orders should be sent with payment to the Welding Research Council, 345 E. 47th St., New York, NY 10017.

228-s|JULY 1982